BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:America/Denver
X-LIC-LOCATION:America/Denver
BEGIN:DAYLIGHT
TZOFFSETFROM:-0700
TZOFFSETTO:-0600
TZNAME:MDT
DTSTART:19700308T020000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=2SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:-0600
TZOFFSETTO:-0700
TZNAME:MST
DTSTART:19701101T020000
RRULE:FREQ=YEARLY;BYMONTH=11;BYDAY=1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20260422T000604Z
LOCATION:E Concourse
DTSTART;TZID=America/Denver:20231115T100000
DTEND;TZID=America/Denver:20231115T170000
UID:submissions.supercomputing.org_SC23_sess301_drs115@linklings.com
SUMMARY:Interactive In-Situ Visualization of Large Distributed Volume Data
DESCRIPTION:Aryaman Gupta (Technical University Dresden, Center for System
 s Biology Dresden (CSBD))\n\nLarge distributed volume data are routinely p
 roduced in numerical simulations and experiments. In-situ visualization, t
 he visualization of simulation or experiment data as it is generated, enab
 les simulation steering and experiment control, which helps scientists gai
 n an intuitive understanding of the studied phenomena. Such data explorati
 on requires interactive visualization with smooth viewpoint changes and zo
 oming to convey depth perception and spatial understanding. As data sizes 
 increase, this becomes increasingly challenging. \n\nThis thesis presents 
 an end-to-end solution for interactive in-situ visualization on distribute
 d computers based on novel extensions to the Volumetric Depth Image (VDI) 
 representation. VDIs are view-dependent, compact representations of volume
  data that can be rendered faster than the original data.\n\nWe propose th
 e first algorithm to generate VDIs on distributed 3D data, using sort-last
  parallel compositing to scale to large data sizes. Scalability is achieve
 d by a novel compact in-memory representation of VDIs that exploits sparsi
 ty and optimizes performance. We also propose a low-latency architecture f
 or sharing data and hardware resources with a running simulation. The resu
 lting VDI is streamed for remote interactive visualization.\n\nWe provide 
 a novel raycasting algorithm for rendering streamed VDIs, significantly ou
 tperforming existing solutions. We exploit properties of perspective proje
 ction to minimize calculations in the GPU kernel and leverage spatial smoo
 thness in the data to minimize memory accesses.\n\nThe quality and perform
 ance of the approach are evaluated on multiple datasets, showing that the 
 approach outperforms state-of-the-art techniques for visualizing large dis
 tributed volume data. The contributions are implemented as extensions to e
 stablished open-source tools.\n\nTag: Accelerators, Artificial Intelligenc
 e/Machine Learning, Applications, Cloud Computing, Distributed Computing, 
 Data Analysis, Visualization, and Storage, Data Compression, Heterogeneous
  Computing, I/O and File Systems, Quantum Computing, Reproducibility, Secu
 rity, Software Engineering\n\nRegistration Category: Tech Program Reg Pass
 , Exhibits Reg Pass\n\n
END:VEVENT
END:VCALENDAR